Hi,
I will send this to the mailing list rather than commenting in the
task as discussions are better on the list (everyone can follow or
even join in). I have responded below:
2009/12/9 Michael Bieri <NO-REPLY.INVALID-ADDRESS@xxxxxxx>:
Follow-up Comment #7, task #6847 (project relax):
Hi Edward
I am stucked in trying to run the full analysis using the main commands. Can
you please help me getting this running? It is the calc_model_free.py
script.
At the moment I am unfortunately extremely busy and will probably only
have a chance after returning from holidays on the 4th of January.
But I will help you will all parts of this development. To get this
running, do you have a simple set of instructions with data that will
show the issues? With a test data set, getting this to work will be
relatively easy. I will also need the most up to date code, so we
need to get all changes into SVN as soon as possible.
I have restructured the whole program (split files, created new ones....).
So
I am not sure if it already makes sense to use the svn yet until there is
something like a structure (which is where I am almost...).
This restructuring will have to go into the repository. This
information is quite important and as the code is now within the
repository there is no choice. All I need is the following
information written out in a post:
Split files:
file1.py:
-> file2.py
-> file3.py
or:
file1.py:
-> file1.py
-> file2.py
Renamed files: file1.py -> file2.py
New files: gui_bieri/res/file5.py
Directory renames: gui_bieir/res -> gui_bieri/model_free
Deleted file: file8.py
I can then put all this information into the repository. A reason for
the change will be needed for a commit message. In preparation for
granting you full commit access to the repository, it would be good if
you wrote the message yourself (I can then give suggestions to help
you learn this coding technique). You can then type:
$ svn up
to get these changes. I would do this on a copy though rather than on
the one you are modifying just in case there are problems. After I
have done this, the changes can then be submitted as patches with
commit messages. This might seem a little painful, but this is how
software is developed. All history is stored in the repository and
the changes are in the form of many small commits. If developed in
this way (the only way possible in a group project), then all relax
developers have instant access to the code to help review it, help
resolve issues, or even write code themselves.
I would be very happy if you (or somebody else) could solve the problems to
run model-free calculations. The rest is working.
I can help, but we need to get the newest changes into the repository.
The patch you attached is 3000 lines long, which is too big for me to
review and apply. It is also missing all the file structure changes
which need to be put into the repository. Can the changes be broken
up into logical blocks? I.e. the change from diropenbox() to
opendir(), is the reason for that the same as the change from "if
len(t2_list3) == 2" to "if len(t2_list3) >= 2"? Each reason for a
change needs to be one patch with one commit message. Note that the
massive loss of lines from gui_bieri/relax_gui.py would, with the
subversion file structure tracking, be accompanied by a whole series
of line additions to another file. I can help you getting this
working, but I will need help as well.
Cheers,
Edward
Re: [task #6847] The Bieri graphical user interface.